Finding that needle: searching tips

I’ve spent the last few months constructing ready made searches for Nutrition and Food Sciences on CAB Direct. Here are some general handy tips for choosing the words for searches that I have gathered on the way. This isn’t the CAB Direct helpfile, you’ll find that on CAB Direct.

What do you want? It’s always a good idea to start by writing down what subjects you want to find and to identify what the concepts are behind them. For example: I want records for articles published in 2006 about B vitamin supplements for preventing heart disease. The concepts, often referred to as ‘keywords’, when searching, are B vitamins, supplements, prevention and heart disease.

1. Know your thesaurus

The first question you should ask yourself is: does the database you are searching contain a thesaurus? These are lists of words that the indexers creating the database use to classify a database record.

If a thesaurus is available, use it to discover the terms indexers will add to the records that interest you. Include these terms in your search keywords.

At CABI, we use a thesaurus to classify records. Its worth knowing that we index the main concepts of a paper, not every keyword. Some CABI thesaurus terms for my example are: vitamin B complex, folic acid, pyridoxine, vitamin B12, and riboflavin.

In CAB Direct, thesaurus terms are added to the descriptors fields. There are 3 descriptor fields: one for taxonomic terms, one for geographical terms and one for all the other keywords. Broader terms are generated from geographical terms and taxonomic terms for plants and animals. These are added automatically to the Broad Term field (BT).

The next question to ask is: Can you ‘Explode’ thesaurus terms to include narrower terms? In Thesauri lists of related terms are arranged in a tree-like structure, with words that are more specific (narrower) grouped under more general terms (broader). ‘Exploding’ a term means the more specific terms are also searched automatically. If there is not an ‘Explode’ option you may need to include narrower terms in your search by typing them in or selecting them from a drop down list. In my example ‘folic acid’ is a narrower term of ‘vitamin B complex’ and I would include it in my search.

Now you have a choice to make: descriptors only or ‘free text’?

Using thesaurus terms and restricting your search to the descriptors fields, will retrieve the records in which your term is an important concept. You will also retrieve fewer records, where your keyword is mentioned in the abstract but is not important to that study. Searching without any field restriction (known as ‘Free Text’ searching) will give you all the records (maximum retrieval) containing the word in your search but you will find more irrelevant records. You need to decide what you want.

Whatever you decide however, you need to use any relevant thesaurus terms in your search.

2. Are there several words for the same concept?

If there are several possible words for the single concept that you want to search on, it may be worth including them all in your search, especially if you are doing a free text search or if there isn’t a thesaurus term to use. For example, for vitamin B12 is also known as cobalamin, and riboflavin is vitamin B2. CABI often puts such words in the Identifier field if they are listed in the thesaurus as the non-preferred term.

You can search the thesaurus fields and the identifier field together if you choose the ‘Subject Term’ field in your search. This can be selected from the drop down menu in the CAB Direct Advanced Search mode, or by using the field tag SU (e.g., riboflavin:SU)

Also remember that there are English/American spelling variations to take account of as well as plurals. Include all these in any free text search. You can often use truncations and ‘wild card’ characters to allow for these (e.g. using vitamin* as your keyword searches for the words ‘vitamin’ or ‘vitamins’ on CAB Direct).

3. Is your term ambiguous?

Got a keyword meaning two things? The term ‘Lens’ is an example of this problem; the Latin name for lentil plants starts ‘Lens’ but of course there are eye lenses as well. A free text search using the keyword ‘lens’ could pick up records on lentils and on eye lenses. Using the relevant thesaurus term in your search and restricting to descriptor fields can help to reduce the number of irrelevant records for ambiguous words. In the ‘lens’ example using the thesaurus term ‘lens’ which is specifically used for lentil plants you will filter out the papers on eye lenses.

You can also get spurious results from data in non-abstract fields. I was searching on food colourants recently and found it was no good using free text searching because some of the colourants that I wanted to retrieve resembled people’s names – e.g. ‘Green S’ and I was getting hits from the Authors names field. This is a situation where restricting the fields that you search may be helpful, rather than doing a search on all the fields. In this case I restricted the search to the abstract field only.

Looking at what you get back from a search will show you if you have used something ambiguous. Looking at the thesaurus helps too!

Questions?

Got a request for a search on Nutrition and Food Sciences? Email me (i.hoskins@cabi.org) and it could be the next prepared search on the Nutrition and Food Sciences Web site. For more general search help or to enquire about other user guides or informal, Web-based training, please contact our International Training Manager, Chris Ison (c.ison@cabi.org).

One thought on “Finding that needle: searching tips

  1. Sarbojit May 12, 2008 / 11:11 am

    Nice site, please give some tme to visit and I’m sure you won’t be disappointed

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s